skills.rest
stefanfaur
stefanfaurCOMMUNITY
Explore all skills by this author →
GitHub

researching-codebase

Documents existing codebase behavior through parallel sub-agent research into structured markdown reports.

1|Updated Feb 10, 2026
Search Tags:#architecture#code analysis#documentation#markdown#github#codebase research#sub-agents
One-click install
npx skills add https://github.com/stefanfaur/roach-marketplace --skill researching-codebase-stefanfaur
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: researching-codebase
Source: https://github.com/stefanfaur/roach-marketplace/tree/main/roach/skills/researching-codebase
Command: npx skills add https://github.com/stefanfaur/roach-marketplace --skill researching-codebase-stefanfaur

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? Answering questions about how a large codebase works requires manually tracing files, components, and interactions, which is slow and produces undocumented, unshareable findings. ## Core Features & Use Cases - Parallel Sub-Agent Research: Spawns specialized locator, analyzer, and pattern-finder agents concurrently to map where code lives and how it works. - Structured Research Documents: Generates dated, domain-scoped markdown reports in thoughts/shared/research/ with YAML frontmatter, code references, and architecture documentation. - GitHub Permalinks & Follow-ups: Converts file references into commit-pinned GitHub links and appends follow-up research to the same document with updated metadata. - Use Case: Ask how the authentication flow works, and receive a synthesized document describing the exact files, line numbers, and component interactions as they exist today. ## Quick Start Ask the agent to research how a specific feature or component works in the current codebase and save the findings as a research document.

Frequently Asked Questions about researching-codebase

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I research how a codebase works with AI agents?▼

Provide a research question, and the skill decomposes it into parallel sub-agent tasks using locator, analyzer, and pattern-finder agents. Findings are synthesized into a structured markdown document with file paths and line numbers.

Where are codebase research documents saved?▼

Research documents are saved under thoughts/shared/research/<domain>/ with filenames in YYYY-MM-DD-description.md format. Each document includes YAML frontmatter with date, researcher, git commit, branch, and tags.

Does the codebase research skill suggest improvements or refactoring?▼

No. The skill strictly documents the codebase as it exists today and explicitly prohibits recommendations, critiques, root cause analysis, or refactoring suggestions unless the user explicitly requests them.

Can I add follow-up questions to an existing research document?▼

Yes. Follow-up questions append a new timestamped section to the same document, and the frontmatter fields last_updated, last_updated_by, and last_updated_note are updated accordingly.

Does the research report include GitHub links to code?▼

Yes, when on the main branch or a pushed commit, the skill generates GitHub permalinks using gh repo view and replaces local file references with commit-pinned URLs including line anchors.

Related Skills

Explore More Software Engineering Skills→
mpsuessermpsuesser

librarian

Research open-source library internals and generate GitHub permalinks for code evidence.

Community
Advanced
idealic-aiidealic-ai

document

Aligns project documentation with code changes through a phase-driven workflow.

Official
Advanced
visualityplvisualitypl
4

research_codebase

Document existing codebases by spawning parallel sub-agents for component analysis.

Official
Advanced
Jamie-CuiJamie-Cui
28

research-explore

Decompose research questions, delegate sub-tasks to an explore agent, and synthesize findings into a structured report.

Community
Advanced